>>>>> On Mon, 18 Jun 2018, Mart Raudsepp wrote:

> Ühel kenal päeval, E, 18.06.2018 kell 08:01, kirjutas Ulrich Mueller:
>> > > > > > On Sun, 17 Jun 2018, Marty E Plummer wrote:
>> > Signed-off-by: Marty E. Plummer <[email protected]>
>> 
>> Please don't yet use Signed-off-by. Our policy on this (GLEP 76) is
>> still a draft and subject to change.
>> 
>> > -DEPEND=">=sys-apps/sed-4"
>> > -
>> > +# sed needs to be executable on the build system
>> > +BDEPEND=">=sys-apps/sed-4"
>> > +[[ ${EAPI:-0} == [0123456] ]] && DEPEND="${BDEPEND}"
>> 
>> These >=sed-4 dependencies are a relic from the time when there
>> was a sed-3. It left the tree in 2003 though.
>> 
>> So I suggest to remove this build-time dependency altogether.

> I prefer being explicit, but that's probably a lost "battle" for sed,
> so, OK, we can remove it instead.

sed is guaranteed to be available in the ebuild environment (as well
as bash, patch, and find):
https://projects.gentoo.org/pms/7/pms.html#x1-11800012.1.1

So in this special case the dependency is entirely redundant.

> Otherwise as a generic question for EAPI-7 support in other eclasses:
> Doesn't this case leak a BDEPEND variable to ebuilds for lower EAPIs,
> as there's no "unset BDEPEND" for the lower EAPIs. And is that OK?

IMHO it is not OK. Also "unset BDEPEND" in global scope is not
entirely clean, because BDEPEND is just a normal variable in older
EAPIs, and could be used by an ebuild.

For the eclasses that I have updated, I have done something along
these lines:

    DEPEND="cat-foo/bar"
    case ${EAPI:-0} in
        0|1|2|3|4|5|6) ;;
        *) BDEPEND="${DEPEND}"; DEPEND="" ;;
    esac

It simply assigns DEPEND in older EAPIs, and in EAPI 7 it assigns
BDEPEND. The additional DEPEND="" assignment should not harm, because
DEPEND is treated specially by the package manager and accumulated
across eclasses.

Ulrich

Attachment: pgpvkEpdOyrqA.pgp
Description: PGP signature

Reply via email to